Skip to main content

Full text of "Data Management In Cellular Networks Using Activity Mining"

See other formats


ACEEE Int. J. on Information Technology, Vol. 01, No. 02, Sep 201 1 



Data Management In Cellular Networks Using 

Activity Mining 



Dr. R. Ramachandran, Ph.D. 

Principal 

Dept. of CSE,SVCE 

Chennai, India 

rrama@svce.ac.in 

V. Aishwaryalakshmi 

PG Scholar 

Dept. of CSE, SVCE 

Chennai, India 

neef avenkat @ gmail . com 



Abstract — In the recent technology advances, an increasing 
number of users are accessing various information systems 
via wireless communication. The majority users in a mobile 
environment are moving and accessing wireless services for 
the activities they are currently unavailable inside. We 
propose the idea of complex activity for characterizing the 
continuously changing complex behavior patterns of mobile 
users. For the purpose of data management, a complex activity 
is copy as a sequence of location movement, service requests, 
the coincidence of location and service, or the interleaving of 
all above. An activity may be composed of sub activities. 
Different activities may exhibit dependencies that affect user 
behaviors. We argue that the complex activity concept provides 
a more specific, rich, and detail description of user behavioral 
patterns which are very useful for data management in mobile 
environments. Correct exploration of user activities has the 
possible of providing much higher quality and personalized 
services to individual user at the right place on the right time. 
We, therefore, propose new methods for complex activity 
mining, incremental maintenance, online detection and 
proactive data management based on user activities. In 
particular, we develop pre-fetching and pushing techniques 
with cost sensitive control to make easy analytical data 
allocation. First round implementation and simulation results 
shows that the proposed framework and techniques can 
significantly increase local availability, conserve execution 
cost, reduce response time, and improve cache utilization. 

Keywords- Acti\\ty mining, Prefectching„pushing, proactive 
data management, Genetic Algorithm; 

I. Introduction 

Diverse mobile services and development in wireless 
networks have stimulated an enormous number of people to 
employ mobile devices such as cellular phones and portable 
laptops as their communications means. The most salient feature 
of wireless networks is mobility support, which enables mobile 
users to communicate with others regardless of location. 
Majority of users in a mobile environment do not travel at random. 
They navigate from place to place with specific purposes in 
mind. In many cases, the patterns of location movement and 
service invocation of mobile users targeting similar purposes 
show strong similarity. The common patterns of location 
movement may due to geographic relationships between 

©2011 ACEEE 
DOI:01.DIT01.02.512 



38 



locations or service distribution. The regularity in service 
invocation may come from the dependencies between services 
or the proximity of service providers. It is potentially beneficial 
to discover such mobility and service patterns to make easy 
network and data management. We propose the idea of complex 
activity for characterizing the continuously changing complex 
behavior patterns of mobile users. A complex activity is a 
sequence of location movement, service requests, the 
coincidence of location and service, or the interleaving of all 
above. An activity may be composed of sub activities. Different 
activities may show dependencies that affect user behaviors. 
We argue that the activity concept provides a more specific, 
rich, and detail description of user behavioral patterns which 
are very useful for data management in mobile environments. 
Proper exploration of such activities enables data management 
system to predict the user's next move, intended service or 
both for providing much higher quality and personalized 
services to individual user at the right place on the right time. 
Such kind of advanced information services call for new 
methods of complex activity mining, incremental maintenance, 
online detection, and proactive data management based on 
user activities. We propose new pattern mining and patterns 
processing algorithms to the discovery and maintenance of 
complex activities in mobile environments. Furthermore, we 
develop pre-fetching, pushing, and handoff techniques with 
cost sensitive control to make easy predictive data allocation 
and personalized services. Preliminary implementation and 
simulation results demonstrate that the activity-based 
proactive data management strategies can significantly 
conserve execution cost, reduce response time, improve cache 
utilization, and increase local availability. 

II. RELATED WORK 

It has been known for quite sometime that user behavior 
patterns are important for effective mobile computing. The 
First stage along this line of research is the acquisition of 
user behavior. Among the various methods for learning user 
behavior, data mining is probably the most widely used 
technique. It is well suited for discovering hidden patterns 
from large volume of data such as transaction log. The mining 
of mobility patterns, in particular focus attention of some 
previously proposed work. 



^ACEEE 



ACEEE Int. J. on Information Technology, Vol. 01, No. 02, Sep 201 1 



• User Mobility Profile (UMP) is a combination of historic 
records and predictive patterns of mobile terminals, which 
serves as fundamental information for mobility management 
and enhancement of Quality of Service (QoS) in wireless 
multimedia networks. UMP framework is developed for 
estimating service patterns and tracking mobile users, 
including descriptions of location, mobility, and service 
requirements. For each mobile user, the service requirement 
is estimated using a mean square error method. Moreover, a 
new mobility model is designed to characterize not only 
stochastic behaviors, but historical records and predictive 
future locations of mobile users as well. Therefore, it 
incorporates aggregate history and current system parameters 
to acquire UMP. In particular, an adaptive algorithm is 
designed to predict the future positions of mobile terminals 
in terms of location probabilities based on moving directions 
and residence time in a cell. The authors G. Resta and P. 
Santi,.[l] have proposed their schemes about User Behavior 
model approach in mobile environments and I.F. Akyildiz 
and W. Wang, [2] describes "The Predictive User Mobility 
Profile Framework for Wireless environments. 

• R.V. Mathivaruni and V.Vaidehi[3] proposes An activity 
based mobility prediction strategy using markv modeling for 
wireless networks which describes the foremost objective 
of a wireless network is to facilitate the communication of 
mobile users regardless of their point of attachment to the 
network. The system must discern the location of the mobile 
terminal, to afford flawless service to the mobile terminal. 
Mobility prediction is widely used to assist handoff 
management, resource reservation and service pre 
configuration. Prediction techniques that are currently used 
don't consider the motivation behind the movement of mobile 
nodes and incur huge overheads to manage and manipulate 
the information required to make predictions. This paper 
proposes an activity based mobility prediction technique that 
uses activity prediction and Markov modeling techniques to 
devise a prediction methodology that could make accurate 
predictions than existing techniques. 

• On The effect of group mobility to data replication in Ad- 
hoc networks: In this paper, we address the problem of replica 
allocation in a mobile ad-hoc network by exploring group 
mobility. We first analyze the group mobility model and derive 
several theoretical results. In light of these results, we propose 
a replica allocation scheme to improve the data accessibility. 

• A. Yamasaki, H. Yamaguchi, S. Kusumoto, andT Higashino 
[4] design a Mobility-aware Data management (MoDA) 
scheme for mobile ad hoc networks (MANETs) composed by 
mobile nodes such as urban pedestrians and vehicles. By 
fully utilizing the knowledge about the trajectories of mobile 
nodes, MoDA determines how replicas of data are copied 
and transferred among mobile nodes to provide the required 
data accessibility. Experimental results have shown that 
MoDA could achieve the small number of data transfers 
among mobile nodes while keeping reasonable accessibility. 



©2011 ACEEE 
DOI:01.DIT01.02.512 



39 



III. COMPLEX ACTIVITIES 

A Behavior pattern 

We analyze the user behaviour patterns in Mobile 
environments and formally characterize the idea of complex 
activities. Users in a mobile environment travel from one 
place to another. The moving pattern of a user can naturally 
be modelled as sequences of locations visited by the user. 
Mobile users also invoke services one after another when 
travelling. Service patterns emerge if a sequence of services 
is repeatedly invoked by the same or different user. Some 
services are offered only at specific locations such as gas 
stations, restaurants, movie theatres, etc. In such cases, 
locations and service invocations always come in pairs. Based 
on the Analysis above, we can characterize the basic user 
behaviour Patterns into three categories: 
Location-only patterns (L-type): Sequences of locations 
Those are repeatedly visited by mobile users. 
Service-only patterns (S-type): Sequences of services Those 
are repeatedly invoked by mobile users. 
Location-service patterns (LS-type): Sequences of Location- 
service pairs that are repeatedly visited and invoked by mobile 
users. Location-only patterns occur when users are engaged 
in Movement activities that are formed by geographical 
Constraints. Service-only patterns arise when users are 
embarked on invocation activities that are formed by service 
dependencies. Location-service patterns are commonly seen 
when there are strong associations between location and 
Service pairs. Sometimes, patterns are the result of personal 
habits. These patterns constitute the primitive activities in 
Mobile environments. Built on top of primitive activities, a 
complex activity can be a combination of any number of 
primitive or other complex activities as sub activities. Complex 
activities: A complex activity A ={al,a2.. } is a frequently 
occurring sequence of activities such That each ai is either a 
primitive or a complex activity. Based on the recursive 
definition, a primitive activity is Considered as the simplest 
type of complex activity. Therefore, the term activity is used 
when there is no need to Distinguish between them.. We 
argue that the notion of Complex activities hold the key to 
effective data management in mobile environments for several 
reasons: Complex activities can be used to faithfully model 
User behaviour in mobile environments. By identifying the 
current activity of a mobile user, we can predict his/her next 
move on both location and service invocation with much 
higher accuracy than traditional way of relying on motion 
estimation and accumulated read/write statistics. Once the 
user behaviour can be predicted with high Accuracy, it 
becomes possible to provide proactive Services for the user. 
In other words, we can allocate the required data items and 
reserve the necessary Resources at the right place on the 
right time. 

B. Service Architecture 

Activity mining is the characterizing of continuously 
changing complex behavior patterns of mobile users. Our 
algorithm for activity mining is based on the popular Apriori 
algorithm to identify all primitive and complex activities of a 
user behavior. The activities identified by the mining algorithm 



iVACEEE 



ACEEE Int. J. on Information Technology, Vol. 01, No. 02, Sep 201 1 



do not have any structure information except for sequences. 
The relationships between activities are not clear. They are 
not well suited for activity processing and data management. 
For this purpose we devise an efficient data structure named 
as activity tree for the representation and incremental 
maintenance of activities. After successful identification of 
activity patterns, we still need to recognize the current activity 
of a user. Once the activity of a user can be detected and 
prioritized, we are all set for presenting the proactive data 
management techniques namely 

(a) Proactive pushing 

(b) Predictive handoff 

(c) Precision pre fetching 

Once the user behavior can be predicted with high accuracy, 

it becomes possible to provide Proactive services for the 

user. In other words, we can allocate the required data items 

and reserve the necessary resources at the right place on the 

right time. To facilitate such an activity-based proactive data 

management services, several issues must be properly 

addressed. These issues, in turn, pose significant challenges 

to information system designers and service providers. 

Activity mining: We must be able to unearth the complex 

activities from large volumes of user behavior logs. This 

naturally demands effective algorithms for data analysis and 

activity mining. 

Activity representation: Once the activities are identified, 

we need efficient structures for representing the activities to 

facilitate effective data management. 

Activity maintenance: Due to highly transient nature of mobile 

environments, complex activities evolve dynamically over 

time. This calls for incremental maintenance techniques that 

can smoothly adapt to the changes. 

Activity detection: After successful identification of activity 

patterns, we still need to recognize the current activity of a 

user. The online detection techniques must be very efficient. 

Otherwise, we may run the risk of late identification of an 

activity that a user is no longer engaged i 

Proactive data management: To provide effective information 

services, proactive or predictive data management techniques 

are highly desirable with the knowledge of a user's possible 

next move or service invocation. The algorithm used is given 

below: 

Function apriori_ gen (Ak_l) 

Ck=0; 

forp2Ak_ldo 

forq2Ak_ldo 

if p:iteml =q:iteml A . . . A p:itemk_2 = q:itemk_2 A 

p:itemk_l "=q:itemk_l then 

Ckp=q:itemk_l; 

end 

end 

for c 2 Ck do 

for {k "l]-subsets s of c do 

if s € Ak_l then delete c from Ck; 

end 

end 

return Ck; 

Function contain Activity (t, c) 

©2011 ACEEE 
DOI:01.DIT01.02.512 



Input: t= {al; a2; . . . ; an} is a transaction and 

c= {cl; c2; . . . ; cm} is an activity candidate. 

if c ==0 ; then return True; 

if n < m then return False; 

for (i = 1 ; i d"n -m, 1++) do 

if action Activity Match { ai; c 1 } A 

contain Activity} aipl; ... ; an}, }c2; . . . ; cm} then 

return True; 

end 

Return False; 

Function action Activity Match (a, c) 

Input: a is an action and c is a primitive activity. 

If c:location '" null A c:service'" null then 

Return a:location ==c:location A a:service == c:service; 

else if c:location "' null then 

return a:location == evocation; 

else 

return a: service == c: service; 

C. Proactive Datamangement 

Proactive data management can contain following steps 

• Activity detection 

• Activity weighting and ranking 

• Data management strategies 

• Activity maintenance 

In the Activity Detection, User maintains the activity table 
that store the current move of the user from the activity tree 
index. In the Activity Weighting and Ranking, Ranking the 
possible move of the user by use of 

1. Service distance(d) 

2. Service data size(t) 

3. Intensity(p) 

4. Conformance ratio(y) 

5. Degree of sharing(w) 

6. Minimal co occurrence(m) 

In the data management strategies, we can find how to data 
transfer and maintenance, that is If cell provide L only- 
proactive pushing. Else If cell provide S only- Predictive 
handoff else If cell provide L & S-Precision pre fetching 
In the Activity Maintenance, they Update the behavior details 
in the behavior table. For an activity a and the set of sharing 
nodes W, we define the activity score of a as follows: 

a.score=(dxt)("ai(X(ai)xa(parent(ai))m))/W 
Once the activity of a user can be detected and prioritized, 
we are all set for presenting the proactive data management 
techniques. In general, when a user is recognized to be 
engaged in an activity, we can predict with high probability 
the user's next move (including location and/or service 
invocation). 

Proactive pushing — When the next move of a user is a 
service-only activity, we know what the user is likely to invoke 
but don't know where he/she is heading. Therefore, it is 
potentially beneficial to push the service data directly to the 
client cache such that it is immediately available when the 
service is actually invoked. 

Predictive handoff — When the next step is a location-only 
activity, we know where the user is going with no knowledge 
of the service invocation. 



40 



4cACEEE 



ACEEE Int. J. on Information Technology, Vol. 01, No. 02, Sep 201 1 



Precision pre fetching — If the next activity is a location 
service pair, we know exactly the user's next move and what 
service the user is going to invoke.. With the strategies in 
mind, selecting proper data management operations becomes 
straightforward. For a user engaged in a complex activity, we 
check the activity tree to retrieve the next activity. If it is a 
service-only activity, we proactively push the service data 
toward the user. If it is a location-only activity, we contact 
the base station of the predicted location and send out the 
data used by the current service. Finally, if it is a location- 
service pair, we inform the base station of the next location to 
initiate the pre fetching of the predicted service data. 

D Cost sensitive Operation 

Even with careful activity identification and proper data 
management operation selection, it may not be always cost 
effective to apply the selected operations. The purpose of 
cost-sensitive operation control is to judiciously estimate 
the relative cost and benefit before applying the operations. 
They are only performed when the expected saving is higher 
than the management overhead. For proactive pushing, we 
push the predicted service data directly to the client cache. 
The pushing cost is, therefore, the service data transfer cost 
from the source to the client as follows: S(fd+1) If the 
prediction is correct, the user's request can be satisfied 
immediately without further delay or cost. However, with 
probability l"a, we may need the extra cost of waiting and 
transferring the data of the service actually invoked by the 
user when the prediction is wrong. This can be characterized 
as follows: (l-a)((D+R/Bf d+R/Bt)l+R(fd+l) 



n 



mechanism that, based on the rewards received, select which 
strategy should be applied at the given moment of the search. 
Suppose we have K > 1 strategies in the pool A = {al, • • • , 



D + T 



Of 



Since a can be estimated by the conformance rate while C, D, 
and T can also be measured once the service data are 
identified, can be used for controlling the application of the 
proactive pushing strategy. Degree of sharing, a data 
management operation can benefit many users at the same 
time which provides a form of consideration on overall system. 
Thus system intergration. 

IV. PROPOSED WORK 

A Adaptive Genetic Selection Metastatergies: 

Met strategies for Adaptation that is quickly adjust the 
system when the situation changes, by making use of 
Adaptive generic algorithm to predict user next move further 
accurately. One more inclusion is activity prioritization that 
we can provide first priority to more important service than 
others in the complex user behavior. On the Adaptive Strategy 
Selection (AdapSS), in the Genetic Algorithms community, 
its objective is to automatically select between the available 
possibly ill-known mutation strategies, according to their 
performance on the current search/optimization process. To 
do so, there is the need for two components: the credit 
assignment, that defines how the impact of the strategies on 
the search should be assessed and transformed into a 
numerical reward; and the strategy (or operator) selection 

©2011 ACEEE 
rX)I:01.DIT01.02.512 



a } and a probability vector P(t) = { pl(t), 



■ P K «} 



In this work, the PM (Probability Matching) technique is 
used to adaptively update the probability pa (t) of each 
strategy a based on its known performance (frequently 
updated by the rewards received). Denote ra (t) as the reward 
that a strategy a receives after its application at time t. qa (t) 
is the known quality (or empirical estimate) of a strategy a, 
that is updated as follows 

qa(t + 1) = qa(t) + a[ra(t) . qa(t)] 

Where a £ (0. 1] is the adaptation rate. Based on this quality 
estimate, the PM method updates the probability pa (t) of 
applying each operator as follows 

Pa(* + 1) =PnHn + (l--K ■PnHa) ! 

Where pmin £: (0, 1) is the minimal probability value of each 
strategy, used to ensure that no operator gets lost. 

1) "DE/randfl": 

v l = x ri + J F-(x r ,-K IS ) (10) 

2) "DEfrandfZ": 

3 J "DEftand Ifrfcesl/T: 

Vi = Xt i +F ■ (xtnt -Xti J+F-fcra-XraJ + F' ()ifr*r,) 

(13) 

it •■DEfaimalMo-rantlM": 

v, = x, + F-(x fb i() + F (x^ Xfi ) CIS) 

B. Combining Road and Ring topology 

Topology information and mobility prediction are main 
distinction between the proposed algorithms and is 
exploitation of road topology map of the destination hotspot 
within a cellular network. Road network topology is 
represented by a stochastic finite state machine that is 
adopted directly from the digital map. Every edge in the map 
is represented by one state ki for each driving direction. 
Road topology is reorganization and management of node 
parameters and modes of operation from time to time to modify 
the network with the goal of extending its Lifetime while 
preserving other important characteristics. Aring network is 
a network topology in which each node connects to exactly 
two other nodes, forming a single continuous pathway for 
signals through each node - a ring. 

C. Prefecting techniques: 

Mobility prediction is an important maneuver that 
determines the location of the mobile terminal by carefully 
manipulating the available information. The prediction 



41 



^cACEEE 



ACEEE Int. J. on Information Technology, Vol. 01, No. 02, Sep 201 1 



accuracy depends on the user movement model and the 
prediction algorithm used. Two different techniques are 
already available for mobility prediction. First technique uses 
the historical movement patterns of the user to predict the 
user's whereabouts in future. The second technique uses 
the contextual knowledge. So, An adaptive algorithm is 
designed to predict the future positions of mobile terminals 
in terms of location probabilities based on moving directions 
and residence time in a cell. It makes use of LAR and DREAM 
routing technique to find the future mobile states of positions 
respectively. 

D. Propsed Architecture .-Algorithm 







T3 



t^.* 



■dl^a M*. *l dlWik* 



Figure 1. Activity mining with metastrategies 

For performance evaluation, we have developed a suite of 
simulation tools using Java language on WinTel platform. 
The layer architecture of the simulator is depicted The 
simulation engine is a set of routines to carry out the 
operations designated by other modules. The mobile 
environment simulator is responsible for generating the 
network topology, supported services, and user behaviours, 
the users get their behaviour models from a behaviour pool. 
Each user has a number of behaviours to choose from at 
random. Each behaviour is associated with a lifetime. When 
the lifetime of all behaviours expires, a new set of behaviours 
is requested from the pool. To keep the behaviours fresh, 
each model has a lifetime in pool to trigger the generation of 
new models. Behaviours are selected fromthe pool following 
Zipf distribution. 

E. Implementation ofZRP andAZRP 

The Adaptive Zone Routing Protocol, as its name implies, 
is based on the concept of zones. A routing zone is defined 
for each node separately, and the zones of neighboring nodes 
overlap. The routing zone has a radius expressed in hops. 
The zone thus includes the nodes, whose distance from the 
node in question is at most n hops, where the routing zone of 
S includes the nodes A-I, but not K. In the illustrations, the 
radius is marked as a circle around the node in question. It 
should however be noted that the zone is defined inhops, 

©2011 ACEEE 
DOI:01.DIT.01.02.512 



42 



not as a physical distance. The nodes of a zone are divided 
into peripheral nodes and interior nodes. Peripheral nodes 
are nodes whose minimum distance to the central node is 
exactly equal to the zone radius. The nodes whose minimum 
distance is less than n are interior nodes. The nodes A-F are 
interior nodes; the nodes G-J are peripheral nodes and the 
node K is outside the routing zone. Note that node H can be 
reached by two paths, one with length 2 and one with length 
3hops. The node is however within the zone, since the 
shortest path is less than or equal to the zone radius. The 
number of nodes in the routing zone can be regulated by 
adjusting the transmission power of the nodes. Lowering the 
power reduces the number of nodes within direct reach and 
vice versa. The number of neighboring nodes should be 
sufficient to provide adequate reach ability and redundancy. 
On the other hand, a too large coverage results in many zone 
members and the update traffic becomes excessive. Further, 
large transmission coverage adds to the probability of local 
contention. AZRP refers to the locally proactive routing 
component as the Adaptive IntrA-zone Routing Protocol 
(AIARP). The globally reactive routing component is named 
Adaptive IntEr-zone Routing Protocol (AIERP). AIERP and 
AIARP are not specific routing protocols. Instead, AIARP is 
a family of limited-depth, proactive link-state routing 
protocols. AIARP maintains routing information for nodes 
that are within the routing zone of the node. Correspondingly, 
AIERP is a family of reactive routing protocols that offer 
enhanced route discovery and route maintenance services 
based on local connectivity monitored by AIARP. The fact 
that the topology of the local zone of each node is known 
can be used to reduce traffic when global route discovery is 
needed. Instead of broadcasting packets, AZRP uses a 
concept called border casting. Border casting utilizes the 
topology information provided by AIARP to direct query 
request to the border of the zone. The border cast packet 
delivery service is provided by the Border cast Resolution 
Protocol (BRP). BRP uses a map of an extended routing zone 
to construct border cast trees for the query packets. 
Alternatively, it uses source routing based on the normal 
routing zone. By employing query control mechanisms, route 
requests can be directed away from areas of the network that 
already have been covered. 

V. Conclusion and future work 

We proposed Personalized activity mining and data 
management are expected to provide Even higher quality 
services then the proposed schemes. Obtaining rich Resource 
station and better equipped client services The cold start 
problem owing to the need for an initial behavior log DB 
remains unsolved. We work on that by dynamic tree structure 
and association mining. The Activity Prioritization methods 
can be improved and enables improved service Quality. 
Different Management strategies used for exploring activities 
in mobile computing. AZRP can be regarded as a routing 
framework rather than as an independent protocol. AZRP 
reduces the traffic amount compared to pure proactive or 
reactive routing. Routes to nodes within the zone are 
immediately available. AZRP makes an extension forZRP 



-fcACEEE 



ACEEE Int. J. on Information Technology, Vol. 01, No. 02, Sep 201 1 



protocol that can adapt well to the complicated network with 
nodes moving non-uniformly. AZRP utilizes the excellent 
performance of the hybrid-driven manner of ZRP and 
simultaneously overcomes the bad adaptability of ZRP which 
assumes each node move uniformly and presets the same 
zone radius. For the mobility of nodes is variable in the 
practical networks, our future work may focus on the change 
of the zone radius aroused by the mobility change of nodes. 
This will be more accordant with the reality. 

References 

[1] G. Resta and P. Santi, "WiQoSM: An Integrated Qos- Aware 
Mobility and User Behavior Model for Wireless Data Networks," 
IEEE Trans. Mobile Computing, vol. 7, no. 2, pp. 187- 198, Feb. 
2008. 

[2] A. Yamasaki, H. Yamaguchi, S. Kusumoto, and T. Higashino, 
"Mobility-Aware Data Management on Mobile Wireless 
Networks," Proc. IEEE 65th Vehicular Technology Conf., pp. 679- 
683, 2007 



[3] J.-L. Huang and M.-S Chen, "On the Effect of Group Mobility 
to Data Replication in Ad-Hoc Networks," IEEE Trans. Mobile 
Computing, vol. 5, no. 5, pp. 492-507, May 2006. 
[4JR.V. Mathivaruni and V.Vaidehi" An Activity Based Mobility 
Prediction Strategy Using Markov Modeling for Wireless Networks" 
Proceedings of the World Congress on Engineering and Computer 
Science 2008 WCECS 2008, October 22 - 24, 2008, San Francisco, 
USA 

[5] I.F Akyildiz and W. Wang, "The Predictive User Mobility 
Profile Framework for Wireless Multimedia Networks," IEEE/ ACM 
Trans. Networking, vol. 12, no. 6, pp. 1021-1035, Dec. 2004. 
[6] I.F. Akyildiz and W. Wang, "The Predictive User Mobility 
Profile Framework for Wireless Multimedia Networks," IEEE/ ACM 
Trans. Networking, vol. 12, no. 6, pp. 1021-1035, Dec. 2004. 
[7] W.-S. Soh and H. Kim, "QoS Provisioning in Cellular Networks 
Based on Mobility Prediction Techniques," IEEE Comm. Magazine, 
vol. 41, no. 1, pp. 86-92, Ian. 2003. 

[8] W. Su, S.-I. Lee, and M. Gerla, "Mobility Prediction and Routing 
in Ad Hoc Wireless Networks," Int'l I. Network Management, vol. 
11, no. l,pp. 3-30, Ian. 2001. 



©2011 ACEEE 
LX)I:01.IJIT01.02.512 



43 



ACEEE